Date anchor: Today's date comes from the # currentDate system context. Compute ALL relative dates from it. Example: if today = 2026-03-14, then "prior month" = 2026-02-01 to 2026-02-28, "current month" (most recent complete) = February 2026, "three months ago" = December 2025. Never use training-data dates.

Depreciation Tracker — Residual Risk Intelligence for Auto Lenders

Lender Profile (Load First)

Load the marketcheck-profile.md project memory file if exists. Extract: state, tracked_segments, portfolio_focus, risk_ltv_threshold, high_risk_ltv_threshold, country. If missing, ask. US-only (requires get_sold_summary + search_active_cars). Confirm profile.

User Context

Lender (residual analyst, portfolio risk manager, auto finance director) tracking collateral depreciation for residual forecasts, advance rates, and portfolio exposure. Also serves lease residual setters and floor plan providers.

Required	Field	Source
Yes	Make/Model (or segment)	Ask
Recommended	Model year(s)	Ask
Auto/Ask	Geography	Profile or ask
Optional	Inventory type	New/Used (default: Used)
Optional	Comparison	EV vs ICE, SUV vs Sedan, Brand A vs Brand B
Optional	Time horizon	30d, 90d, 6mo, 1yr

Clarify: used vehicle depreciation vs new vehicle MSRP parity — different workflows.

Workflow: Make/Model Depreciation Curve

Use this when a user asks "how fast is the RAV4 losing value" or "show me the residual risk curve for a 2022 Civic."

1. Get current period sold data — Call get_sold_summary with make, model, inventory_type=Used, date_from (first of prior month), date_to (end of prior month). Include state if specified. → Extract only: average_sale_price, sold_count. Discard full response.

2. Get historical sold data at multiple intervals — Make separate calls to get_sold_summary for each lookback period:

   • 60 days ago
   • 90 days ago
   • 6 months ago
   • 1 year ago Record average_sale_price at each point. Adjust dates based on today's date. → Extract only: average_sale_price per interval. Discard full response.

3. Get current active market asking price — Call search_active_cars with year, make, model, car_type=used, stats=price, rows=0. Include zip/state if available. → Extract only: mean, median, min, max price stats. Discard full response.

4. Get original MSRP baseline — Call search_active_cars with same YMMT, rows=1, sort_by=price, sort_order=desc. Decode the VIN for MSRP. Fallback: highest 1-year-ago sold price. → Extract only: msrp from decode. Discard full response.

5. Build the residual risk curve — Calculate at each time interval:

   • Retention % = (average_sale_price at interval / original MSRP) x 100
   • Monthly depreciation rate = (price change between consecutive intervals) / (months between intervals)
   • Annualized depreciation rate = monthly rate x 12
   • LTV implication = If the user provided a loan amount, calculate LTV at each interval: (loan balance / current market value) x 100 Present as a table and describe the curve shape (linear, accelerating, stabilizing). Flag any interval where the monthly rate exceeds 2% as "residual risk increasing."

Workflow: Segment Value Trends

Use this when a user asks "are SUVs holding value better than sedans" or "how is EV depreciation compared to ICE."

1. Get current period segment data — Call get_sold_summary with ranking_dimensions=body_type, ranking_measure=average_sale_price, date_from (first of prior month), date_to (end of prior month), inventory_type=Used, top_n=10. → Extract only: per body_type — average_sale_price, sold_count. Discard full response.

2. Get prior period segment data — Same call with dates shifted back 3 months (or user's chosen comparison window). → Extract only: per body_type — average_sale_price, sold_count. Discard full response.

3. Get fuel type comparison — Call get_sold_summary with fuel_type_category=EV, current period dates, inventory_type=Used. Repeat with fuel_type_category=ICE. Repeat both for prior period. → Extract only: average_sale_price, sold_count per fuel_type per period. Discard full response.

4. Calculate segment trends and residual risk signals — For each body type and fuel type:

   • Period-over-period price change = (current avg price - prior avg price) / prior avg price x 100
   • Volume change = (current sold_count - prior sold_count) / prior sold_count x 100
   • Flag segments where price declined more than 3% as "residual risk increasing — tighten advance rates"
   • Flag segments where price held within +/- 1% as "residual risk stable"
   • Flag segments where price increased as "residual risk decreasing — opportunity for competitive rates"

5. Deliver the segment comparison — Present a ranked table from strongest retention to weakest. Highlight the EV vs ICE gap specifically (this is critical for portfolio concentration risk). Include volume context — a segment with strong prices but falling volume may be about to soften, signaling upcoming residual pressure.

Workflow: Brand Residual Ranking

Use this when a user asks "which brands hold value best" or "rank the automakers by residual value."

1. Get current period brand prices — Call get_sold_summary with ranking_dimensions=make, ranking_measure=average_sale_price, ranking_order=desc, date_from (first of prior month), date_to (end of prior month), inventory_type=Used, top_n=25. → Extract only: per make — average_sale_price. Discard full response.

2. Get prior period brand prices — Same call with dates shifted back 6 months (or user's preferred comparison window). → Extract only: per make — average_sale_price. Discard full response.

3. Get volume context — Call get_sold_summary with ranking_dimensions=make, ranking_measure=sold_count, ranking_order=desc, current period dates, inventory_type=Used, top_n=25. → Extract only: per make — sold_count. Discard full response.

4. Calculate brand retention scores — For each make:

   • Retention % = current average_sale_price / prior average_sale_price x 100
   • Volume trend = current sold_count vs prior sold_count (indicates demand strength)
   • Rank brands by retention % descending
   • Separate into tiers: Tier 1 (>98% retention), Tier 2 (95-98%), Tier 3 (90-95%), Tier 4 (<90%)

5. Present the brand ranking with lending implications — Show a ranked table with: Rank, Make, Current Avg Price, Prior Avg Price, Retention %, Volume, Tier, Advance Rate Guidance. For each tier, provide advance rate guidance:

   • Tier 1: Standard advance rates appropriate
   • Tier 2: Monitor — consider 2-3% lower advance rate cap
   • Tier 3: Reduce advance rates 5%+ or require higher down payment
   • Tier 4: High residual risk — restrict advance rates significantly or require GAP coverage

Workflow: Geographic Depreciation Variance

Use this when a user asks "where do Tacomas hold value best" or "which states have the highest used car prices."

1. Get state-level transaction data — Call get_sold_summary with make, model, summary_by=state, date_from (first of prior month), date_to (end of prior month), inventory_type=Used, limit=5000. → Extract only: per state — average_sale_price, sold_count. Discard full response.

2. Get national baseline — Same call without summary_by for national average. → Extract only: average_sale_price, sold_count. Discard full response.

3. Calculate geographic variance — For each state:

   • Price index = state average_sale_price / national average_sale_price x 100 (100 = national average)
   • Premium/discount = state price - national price in dollars
   • Sort by price index descending to show where vehicles command the highest premiums

4. Identify patterns and lending implications — Group states into:

   • Premium markets (index > 105): collateral values stronger, lower LTV risk
   • At-national-average (index 95-105): standard collateral valuation appropriate
   • Discount markets (index < 95): collateral values weaker, higher LTV risk — consider state-level adjustments to advance rates Note regional patterns (e.g., trucks command premiums in mountain/rural states, EVs hold better in CA/Northeast).

5. Deliver the geographic map — Present as a ranked table: State, Avg Transaction Price, National Avg, Price Index, Premium/Discount $, Sold Count. Highlight the top 5 and bottom 5 states for the specific vehicle. Include a recommendation on whether to apply state-level collateral value adjustments.

Workflow: MSRP Parity Tracker

Use this when a user asks "which new cars are selling over sticker" or "are markups coming down" or "incentive effectiveness."

1. Get current MSRP parity data — Call get_sold_summary with inventory_type=New, ranking_dimensions=make,model, ranking_measure=price_over_msrp_percentage, ranking_order=desc, date_from (first of prior month), date_to (end of prior month), top_n=30. → Extract only: per make/model — price_over_msrp_percentage. Discard full response.

2. Get prior period parity data — Same call with dates shifted back 3 months. → Extract only: per make/model — price_over_msrp_percentage. Discard full response.

3. Get volume context — Call get_sold_summary with inventory_type=New, ranking_dimensions=make,model, ranking_measure=sold_count, ranking_order=desc, current period dates, top_n=30. → Extract only: per make/model — sold_count. Discard full response.

4. Classify parity status with lending implications — For each make/model:

   • Above MSRP (price_over_msrp_percentage > 0): strong demand — residual risk lower on these models
   • At MSRP (price_over_msrp_percentage between -1% and 0%): balanced market — standard residual assumptions hold
   • Below MSRP (price_over_msrp_percentage < -1%): incentives/discounts active — residual risk may increase as new vehicle transaction prices drop, pulling used values down
   • Trend direction: compare current vs prior period to show if markups are growing or shrinking

5. Present the parity report — Show a table: Make/Model, Current % Over/Under MSRP, Prior Period %, Change Direction, Sold Volume. Highlight:

   • Models that flipped from above-MSRP to below (lower residual forecasts for in-fleet units)
   • Models still commanding premiums (residual assumptions can be maintained)
   • Models with deepening discounts (reduce residual estimates and tighten advance rates on new originations)

Output

Present: summary headline with residual risk signal, depreciation curve / trend data table(s) with retention %, key risk signals (EV vs ICE gap, LTV implications, geographic variance), and actionable recommendation tied to advance rates or residual forecasts.